Table 3.

Summary of the mixed-effects linear regression on latency to reach reward bucket in successful trials.

Variable	Levels	n	EMM	SE(EMM)	95% CI(EMM)	Df	p
Age	Continuous	29	47.1	3.7	39.5–54.8	1	0.97	ns
Breed	Holstein	18	45.7	4.4	36.7–54.7	1	0.69	ns
	Red	11	48.6	5.9	36.4–60.7
Treatment	Observer	14	41.3	4.9	31.2–51.3	1	0.08	ns
	Control	15	53.0	5.2	42.4–63.7
Trial	First	27	44.5	4.5	35.6–53.5
	Second	19	42.3	5.2	31.9–52.8	2	0.15	ns
	Third	15	54.6	5.9	42.7–66.4
Route	Demonstration	41	39.2	4.2	30.7–47.7	1	0.01	**
	Opposite	20	55.1	5.5	44.0–66.1

Mixed-effects linear regression done with lme4 package. Variables include age (numeric variable), breed (categorical variable with two levels), treatment (categorical variable with two levels), trial (categorical variable with three levels) and route (categorical variable with two levels). The results are on the response scale and estimated marginal means have been calculated with emmeans package for each variable: n, number of observations; EMM, Estimated Marginal Mean; SE_(EMM), standard error of EMM; 95% CI_(EMM), confidence interval of EMM; Df, degrees of freedom and p-values were calculated in ANOVA with car package.

^**p < 0.01. ns, not significant.